智能网格天气预报监控平台设计研究

Application Research of Intelligent Grid Weather Forecast Business Monitoring Platform

近年来,全国气象部门构建了面向“国-省”的智能网格预报业务体系,将传统的站点天气预报转换为精细服务的格点天气预报。但由于业务流程繁杂,且缺少规范化的监控平台,为业务的运行维护带来很大困难。针对此问题,该文分析了当前气象部门各业务领域的监控系统研发现状,结合智能网格预报业务特点,提出将业务流程和数据流程解耦的策略。以数据流为中心,设计了对智能网格天气预报业务“全流程、一体化、可视化”的监控平台。通过采用Kafka消息中间件和Logstash、Cassandra、ElasticSearch、Redis分布式缓存等核心技术,实现了对业务监视数据的建模、采集、处理、存储和展示。目前,智能网格天气预报监控平台已在陕西省气象部门投入使用,运行稳定、可靠性高,有效提高了运维人员的监控效率,保障了智能网格预报业务的顺利开展。

In recent years,the National Meteorological Department has constructed the national-provincial intelligent grid forecast system,which transforms the traditional weather forecast from station to grid weather forecast with fine service.However,due to the complex business process and lack of standardized monitoring platform,it is difficult for the operation and maintenance of the business.In order to solve this problem,we analyze the current situation of monitoring system development in various business areas of the meteorological department.Combined with the characteristics of the intelligent grid forecasting business,the idea of decoupling business process and data process is proposed.Taking data flow as the center,the monitoring platform of"whole process,integration and visualization"for intelligent grid weather forecasting business is designed.The platform has used Kafka message middleware,and core technologies such as Logstash,Cassandra,ElasticSearch,Redis distributed cache to realize the modeling,collection,processing,storage and display of business monitoring data.At present,the platform has been put into use in Shaanxi Meteorological Department with stable operation and high reliability,which effectively improves the monitoring efficiency of operation and maintenance personnel and ensures the smooth development of intelligent grid forecasting business.